In above table, operator choose operating voice in GSM network where UEs that are most likely to use voice service shall camp on GSM layer and hence set GSM as the top priority in priority list; while for packet service like gaming, streaming video, it is preferred to LTE layer thus set LTE as the highest priority.

With these policies, a mapping table could be established for Non-IMS/GSM/LTE case as shown in Table 2. Table 2 Mapping table for Non-IMS/GSM/LTE

RFSP in the table only considers service. In practice however it may comprise some other aspects like Subscriber Class. Thus, according to a step of the method of load balancing according to the invention one would associate, for at least one entry of the mapping table, the Classifier Index (RFSP) additionally with a Subscriber Class, said Subscriber Class discriminating service availability for a subscriber. Thereby, subscribers may be categorized into different classes, each of which is associated with a certain level of mobility/roaming restriction, subscription type etc. For example, define platinum users as without any restrictions, Golden user with few restrictions and normal users with some restrictions. Subscriber Class may also be associated with subscription type, like broadband subscription to be golden user, narrowband subscription to be normal user etc. Therefore according to a step of the method of load balancing according to the invention, said service availability of the Subscriber Class represents any from the group consisting of: number of mobility restrictions, bandwidth availability.

Table 3 shows an example for RFSP taking Subscriber Class into account as well. This categorization undoubtedly influences mobility behavior of UE. The Subscriber Class and service can be encoded together to form RFSP.

Table 3 Mapping table for Non-IMS/GSM/LTE

The RFSP encoding algorithm is then transferred to Home Subscriber Server (HSS) to ensure that RFSP is assigned to a subscriber in a coordinated way with OAM. In Figure 2, RFSP assigning policy is for this purpose.

The distribution policy is set for service by service, in some cases from UE' s history, it is possible that multiple services are of almost same possibility of occurrence, hence it is possible to have one RFSP with more than one most probable service. For example, one RFSP with both voice and gaming are most likely to use as exemplified in table 4. Thus, according to the invention we can associate, for at least one entry of the mapping table, the Classifier Index (RFSP) additionally with at least one additional service of the mobile communications system.

In this case, we need a scheme allowing multiple distribution policies to work together. Table 4 Mapping table for Non-IMS/GSM/LTE

Hence, it is proposed that policy priority is introduced to rank from most important to least important; when multiple policies act simultaneously, the highest priority policy is used.

In some scenarios, operator would like to distribute one service to more than one layer, for example, in

IMS/GSM/UMTS/LTE voice can be actually serviced through any of the three RATs. Then a fractional distribution principle could be set for a certain service. In order to accomplish that, the method of the invention could further comprise a step of associating, for at least one entry of the mapping table, at least one additional priority order list, and for each priority order list assigned to the at least one entry, information on the fraction of UEs, having the Classifier Index of that at least one entry, which should use that specific priority order list. More specifically, fractional distribution can be used for a certain RFSP as shown in Table 5 where 60% of UEs with RFSP 0 (most probable service is voice) will prefer to GSM in camping, 20% of them to UMTS and 20% of them to LTE.

Table 5 Fractional distribution for IMS/GSM/UMTS/LTE

When performing the actual priority assigning to a UE, the entity doing the assigning, e.g. a network access node, has to take into account also the fraction information, as shown in table 5. That is, the method of load balancing according to the invention has to, in the step of assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers, acquire the Classifier Index of the User Equipment and assign, in accordance with the information on the fraction of UEs to use a list, one of the corresponding priority order lists from the mapping table, to the User Equipment as the idle mode priority order of access layers .

As mentioned earlier, capability mapping is either separated from or integrated into service mapping. Correspondingly, mapping table is different. In case of separation, the mapping table performs service mapping only. Capability mapping follows the service mapping without need of mapping table. Here, the capability mapping only considered UE capability because network capability actually has already been taken into account in service mapping. Capability mapping is just a process adapting priority list looked up in the mapping table to UE capability.

In case of one-step process, UE capability is integrated into the mapping table as well. That is, according to the invention at least one entry of the mapping table is associated with at least one UE Capability, which UE Capability defines an ability of a User Equipment (UE) to use an access layer.

Table 6 shows an example for IMS/GSM/UMTS/LTE scenario . Table 6 Ca abilit ma in table for EMS/GSM/UMTS/LTE

As seen in Table 6, for the method of load balancing, at least one UE Capability can be the ability to use any access layer from the group consisting of: GSM, UMTS, LTE. However, other layer types are conceivable, as well as different frequency bands within a RAT.

In this case, when performing the actual priority assigning to a UE, the entity doing the assigning, e.g. a network access node, has to take into account also UE Capability, as shown in table 6. That is, the method of load balancing according to the invention has to, in the step of assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers, acquire the Classifier Index and the UE Capability of the User Equipment and assign the corresponding priority order list from the mapping table, to the User Equipment as the idle mode priority order of access layers .

In the above table, a very simple example of UE capability is used, where each possible terminal type is represented. It may however be necessary to take into account more specific UE capabilities. But this requires a flexible way of defining the UE capability to avoid that the mapping table becomes too large .

Each entry on the UE capability can further be broken down into different classes which we in this invention call UE Capability class. Thus, in the method of load balancing according to the invention, we would further associate with at least one entry of the mapping table, at least one UE Capability Class, which UE Capability Class defines, for a User Equipment (UE) having a UE Capability, a further ability in usage of an access layer.

This solution allows for a flexible definition of UE capability classes where each class can be represented in the mapping table.

According to the invention the at least one UE Capability Class could for instance be the ability to use any access layer feature from the group consisting of: Enhanced Data rates for GSM Evolution (EDGE) , High Speed Packet Access

(HSPA) , HSPA+, Multiple- Input and Multiple-Output (MIMO) , high order modulation. For example, UMTS capable UEs may further be categorized as High Speed Packet Access÷ (HSPA+) capable UEs and non-HSPA+ capable UEs, which can be handled discriminately in practice. This is very useful to distribute UEs that favor special service like gaming which is able to operate on both HSPA+ and LTE. HSPA+ capable UE can be distributed to both UMTS and LTE while non-HSPA+ capable UE may only be distributed to LTE for gaming service.

Table 7 Ca abilit ma in table with UE ca abilit class

UE capability class is defined for each capability mode (uni- /Bi-/Tri-mode) by OAM by specifying a more detailed feature of network under one RAT that UE support. This feature is of course able to support certain services. Like HSPA+ to support gaming service shown in above table. Therefore, when UE capability class is defined, OAM need to take network capability and service association into account.

A further attempt to reduce the size of the table is to allow the RFSP and UE capability class to be defined as a range (for example x-y) instead of single values. Another possibility is to also allow a sequence of values (for example x,y) . This is illustrated in Table 6. Table 6 Ma in table with ran es and se uences

Thus, according to the method of load balancing according to the invention at least one of Classifier Index and UE Capability Class is represented in the mapping table as a range of values.

In order to implement a method step for assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers taking into account UE

Capability class, it is proposed to acquire the Classifier Index and the UE Capability Class of the User Equipment and assigning the corresponding priority order list from the mapping table, to the User Equipment as said idle mode priority order of access layers. This assigning is preferably done by an access network node in the mobile communications system.

The description above assumes that the mapping table is initially built up and subsequently updated with different load situation at OAM and configured to Access network for local mapping. This is a typical approach of using the mapping table to control priority for load balancing where information between OAM and access network are load- related performance indicator and the mapping table as shown in Figure 5. Another approach is that OAM initially configures a set of mapping tables to access network, each of which represents a mapping on a certain load pattern and is associated with a unique table ID. In this case the method of load balancing according to the invention would further comprise to configure at least two different mapping tables and assigning a different Mapping Table ID to each.

OAM signals access network a table ID showing which mapping table it shall use. Whenever OAM, based on its analysis of load-related performance indicator, find necessity of change of mapping table, it reconfigures the corresponding table ID to access network. That is to say, that according to the method of the invention, in response to a change of a load of the mobile communications system a new mapping table could be assigned by signaling from a central node of the mobile communications system, the corresponding mapping table ID to an access network node of the mobile communications system. This approach is very useful to save the signaling between OAM and access network node in the case that the size of the mapping table is considerably large. Figure 3 shows an example of this approach.

In the approach illustrated in Figure 3, the decision-maker on which table shall be used is OAM. In some cases, it is very useful to have access network to make this decision. For example, in a simplified case where unbalanced situation only take local performance indicator into account, e.g. redirection number, idle mode UE number etc., access network is capable of selecting the mapping table based on its analysis of local performance indicator. Thus, according to the method of load balancing according to the invention, there could be a further step comprising, in response to a change of load of the mobile communications system: - using, in an access network node of the mobile communications system, a new mapping table and notify a central node of the mobile communications system of the use of the new mapping table by signaling the corresponding mapping table ID to the central node.

As shown in Figure 4, in this approach, OAM initially configures a set of mapping tables and associated parameters that shall be used for selecting this table to access network. Access network subsequently decide which mapping table shall be used based on its analysis of local load-related performance indicators. Each time the mapping table changes, access network may notify corresponding table ID to OAM.

In the method relating to this it is conceivable to further signal information on a change of load of another Radio Access Technology network to the access network. I.e., in order to further consider the case that cell load join the mapping table selection, cell load statistics of other RATs can be transferred to access network in concern. So in Figure 4, signaling of cell load from OAM to E-UTRAN is kept, otherwise, it is not needed.

The initial mapping table is configured to E-UTRAN and can be updated subsequently when needed. The purpose of mapping table update is to balance load of layers. The first step of mapping table update is detection of unequal traffic of different layers, which in this invention relies on load related performance indicator from access network. OAM collects these performance indicators and evaluate these indicators to decide whether the mapping table update is needed or not. If needed, for example when overload is happening on a RAT layer, OAM will adapt the mapping table according to its algorithm.

Load information is one of the most important indicators used to detect unbalance situation. Load information is reported from various access networks, e.g. GERAN, UTRAN, E-UTRAN as shown in Figure 5, on per cell basis, OAM has the freedom of evaluating them on the granularity of cell, tracking area etc. Each RAT layer is assigned a pre-defined threshold, if the load is above the threshold, overload is regarded happening. So, the load related performance indicator according to the invention can be load information from any access network from the group consisting of GSM EDGE Radio Access Network (GERAN) , UTRAN, E-UTRAN.

Yet another indicator, number of service-caused redirection to/from E-UTRAN, can be used to detect unbalanced situation. Over-threshold number of the redirection to E-UTRAN shows that too many UEs that should have camped on E-UTRAN mismatch their best camping layer. Then in this case E-UTRAN layer's priority shall be lift up to pull these UEs in. On the other hand, if number of redirection from E-UTRAN to other RATs is above a threshold, the priority of E-UTRAN layer needs to reduce down. Number of idle mode UE can also be such an indicator. This indicator can be used to avoid unbalanced number of UE camping even though no overload occurs yet. A variant of this indicator is number of periodic Traffic Area Updates (TAU) .

Another indicator that this invention proposes is load statistics in terms of RFSP of day allowing for traffic mode adaptive mapping table update as will be described in detail later in this application.

To summarize the examples above, the load related performance indicator according to the method of the invention could for instance be any from the group consisting of: number of redirects to or from a radio access network, number of idle mode UEs in a radio access network, number of periodic Traffic Area Updates (TAU) and load statistics in terms of Classifier Index.

RFSP is assigned by CN to distinguish the service history of a UE. It is possible to collect information about how the different RFSP move in the network when they have transited to active mode. If we collect information about which RFSPs are handed over to other systems, and which RFSPs typically stays in the LTE system, it is possible to evaluate if the current strategy is valid or if one RFSP for example has the highest priority set for LTE, but very often is moved to UMTS when entering active mode.

It is also possible to detect cases where the current RFSP assignment in the CN is not giving enough information, for example if 50% of the UE with the same RFSP stays in LTE and the rest is moved to GSM. Then the system can realize that this RFSP is not providing any helpful information and may need to be reconfigured.

If the mapping tables are updated in a central node (centralized architecture) there is no need to coordinate actions between different nodes. If, on the other hand, the decision to change mapping table is decided in the eNB (distributed architecture) there is a need to coordinate the decisions. One way of doing this is to create a policy controlling how to adjust the mapping tables.

For each performance indicator mentioned above, a certain mapping table update policy is needed. As an example, Table 7 shows a set of update policies where load situation, redirection and idle mode UE camping are considered.

Table 7 Mapping table update policy

In other words, the method of load balancing according to the invention could further comprise any of the following steps:

- set, as a mapping table update policy, the priority of an access layer to lowest in a priority order list of at least one entry of the mapping table when load of this access layer is above a threshold. - set, as a mapping table update policy, the priority of an access layer one level up in a priority order list of at least one entry of the mapping table when number of redirections to the layer is above a threshold.

- set, as a mapping table update policy, the priority of an access layer one level down in a priority order list of at least one entry of the mapping table when number of redirections from the layer is above a threshold.

- set, as a mapping table update policy, the priority of an access layer one level up in a priority order list of at least one entry of the mapping table when number of idle UEs camping on the access layer is above a threshold.

In order for all the update policies to be able to work together, a method is needed to ensure that they do not interact negatively with each other. It is proposed in this invention that each update policy is associated with one priority. When multiple policies work together, the highest priority update policy is used. For example, in Table 7, we may set update policy due to load indicator the highest priority, when one layer is overloaded, even though according to redirection policy its priority should be increased, we still put it on the lowest priority.

That is, in the method of load balancing according to the invention we would further be able to prioritize a mapping table update policy such that when two or more policies are true only the one with highest priority is performed.

In some network deployment configurations, the traffic load mode is regularly changed. For example, in the office area, traffic load is high in the daytime but low during the night. The mapping table may be set differently during the day and night .

In order to detect traffic model, some indicator is needed. One indicator proposed in this invention is traffic load in terms of RFSP where total load is the sum of weighted number of RFSPs. Each type of RFSP number is weighted by a weight factor which represent averaged load of each one of this RFSP. i.e.

Total Load=NO_RFSPl*W_RFSPl+NO_RFSP2*W_RFSP2+... Based on observation of this indicator, the network is able to know the traffic change model, and accordingly correlate the percentage of RFSPx and the number of idle mode mobiles in a cell based on the time of day and decide, based on priority of RFSPx, where to push mobiles and which policies to apply.

Hence, in the method of load balancing according to the invention we define, as mentioned above, a total load measure as Total Load = NO_RFSPl*W_RFSPl+NO_RFSP2*W_RFSP2+...+NO_RFSPx*W_RFSPx, wherein N0_RFSPx denotes number of User Equipments in an access layer having a Classifier Index RFSPx and W_RFSPx denotes an average load of all User Equipments having a Classifier Index RFSPx in the access layer.

A specific use of the method of load balancing according to the previous paragraph could be to report a Total Load measure to a central node of the mobile communications system. This would be useful in the case where the central node, for instance an OAM node, is the decision-maker on when to use an updated mapping table due to the variation of the Total Load.

According to the method, then, we would change mapping table in regard of a threshold for a Total Load measure, in order to adapt to the change of load information according to the Total Load Measure. The mapping table change could be initiated in the access network or, as in the previous paragraph, in a central node of the mobile communications system, such as an OAM node .

Figure 6 illustrates steps of a method of load balancing in a mobile communications system according to the invention, the mobile communications system comprising a plurality of access layers. In the uppermost step, it is defined a network deployment configuration comprising information on each different type of access layer of the plurality of access layers . In a second uppermost step of the method, it is defined for the network deployment configuration, a service distribution scheme comprising, for an at least one service of the mobile communications system, an order of preference of the types of the plurality of access layers for the at least one service. In a third uppermost step of the method, the at least one service is associated, for at least one entry of a mapping table, with at least one Classifier Index, in the mapping table, and a to the Classifier index corresponding priority order list is set to be the order of preference of the plurality of access layer types defined in the service distribution scheme. Together, the Classifier Index mapped to the priority order list forms the entry of the mapping table. As is shown in the bottom step of figure 6, the mapping table can be updated in relation to at least one load related performance indicator of the mobile communications system by altering at least one entry of the same. The different steps of the method of the invention described in this application can be combined or performed in any suitable order. A condition for this is, of course, that the requirements of a step, to be used in conjunction with another step of the method of the invention, in terms of, for instance, Classifier Index, Subscriber Class, number of priority order lists, UE Capability Class, etc. must be fulfilled.

The method of the invention can implemented by a computer program, having code means, which when run in a computer causes the computer to execute the steps of the method. The computer program is included in a computer readable medium of a computer program product . The computer readable medium may consist of essentially any memory, such as a ROM (Read-Only Memory) , a PROM (Programmable Read-Only Memory) , an EPROM (Erasable PROM) , a Flash memory, an EEPROM (Electrically Erasable PROM), or a hard disk drive.

Claims

1. Method of load balancing in a mobile communications system comprising a plurality of access layers, the method comprising the steps : - defining at least one Classifier Index (RFSP) classifying User Equipment (UE) of the mobile communications system,

- building a mapping table comprising at least one entry: a Classifier Index mapped to a priority order list for the plurality of access layers, - assigning, to a User Equipment of the mobile communications system, an idle mode priority order of access layers by: acquiring the Classifier Index of the User Equipment and assigning the corresponding priority order list from the mapping table to the User Equipment as said idle mode priority order of access layers, characterized by the following steps of building the mapping table:

- define a network deployment configuration comprising information on each different type of access layer of the plurality of access layers,

2. Method of load balancing according to claim 1, comprising the step: - updating the mapping table by altering at least one entry in relation to at least one load related performance indicator of the mobile communications system.

3. Method of load balancing according to claim 1, wherein the access layers of the plurality of access layers are of any type from the group consisting of: Radio Access Technology (RAT) network, and frequency band of a Radio Access Technology (RAT) network.

4. Method of load balancing according to claim 3, wherein the Radio Access Technology (RAT) network is any from the group consisting of: GSM, UMTS, LTE, WiMAX, CDMA2000.

5. Method of load balancing according to claim 1, wherein in the step of defining at least one network deployment configuration, define said network deployment configuration to also comprising information on capability of a core network of the mobile communications system.

6. Method of load balancing according to claim 5, wherein further defining said capability of a core network to be at least IP Multimedia Subsystem (IMS) .

7. Method of load balancing according to claim 1, further comprising:

- associate, for at least one entry of the mapping table, the Classifier Index (RFSP) additionally with a Subscriber Class, said Subscriber Class discriminating service availability for a subscriber.

8. Method of load balancing according to claim 7, wherein said service availability of the Subscriber Class represents any from the group consisting of: number of mobility restrictions, bandwidth availability.

9. Method of load balancing according to claim 1, further comprising:

- associate, for at least one entry of the mapping table, the Classifier Index (RFSP) additionally with at least one additional service of the mobile communications system.

10. Method of load balancing according to claim 1, further comprising :

- associate, for at least one entry of the mapping table, at least one additional priority order list, and for each priority order list assigned to the at least one entry, information on the fraction of UEs, having the Classifier Index of that at least one entry, which should use that specific priority order list.

11. Method of load balancing according to claim 10, wherein in the step of assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers,

- acquiring the Classifier Index of the User Equipment and assigning, in accordance with the information on the fraction of UEs to use a list, one of the corresponding priority order lists from the mapping table, to the User Equipment as said idle mode priority order of access layers.

12. Method of load balancing according to claim 1, further comprising:

- associate with at least one entry of the mapping table, at least one UE Capability, which UE Capability defines an ability of a User Equipment (UE) to use an access layer.

13. Method of load balancing according to claim 12, wherein the at least one UE Capability is the ability to use any access layer from the group consisting of: GSM, UMTS, LTE.

14. Method of load balancing according to any of claims 12 - 13, wherein in the step of assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers, - acquiring the Classifier Index and the UE Capability of the User Equipment and assigning the corresponding priority order list from the mapping table, to the User Equipment as said idle mode priority order of access layers.

15. Method of load balancing according to claim 13, further comprising:

- associate with at least one entry of the mapping table, at least one UE Capability Class, which UE Capability Class defines, for a User Equipment (UE) having a UE Capability, a further ability in usage of an access layer.

16. Method of load balancing according to claim 15, wherein at least one of Classifier Index and UE Capability Class is represented in the mapping table as a range of values.

17. Method of load balancing according to any of claims 15 - 16, wherein in the step of assigning to a User Equipment of the mobile communications system an idle mode priority order of access layers, - acquiring the Classifier Index and the UE Capability Class of the User Equipment and assigning the corresponding priority order list from the mapping table, to the User Equipment as said idle mode priority order of access layers.

18. Method of load balancing according to claim 1, further comprising:

- configuring at least two different mapping tables and assigning a different Mapping Table ID to each.

19. Method of load balancing according to claim 18, further comprising, in response to a change of a load of the mobile communications system: - assign a new mapping table by signaling from a central node of the mobile communications system, the corresponding mapping table ID to an access network node of the mobile communications system.

20. Method of load balancing according to claim 18, further comprising, in response to a change of load of the mobile communications system:

- using, in an access network node of the mobile communications system, a new mapping table and notify a central node of the mobile communications system of the use of the new mapping table by signaling the corresponding mapping table ID to the central node.

21. Method of load balancing according to claim 20, further comprising,

- signaling information on a change of load of another Radio Access Technology network to the access network.

22. Method of load balancing according to claim 2, wherein the load related performance indicator being load information from any access network from the group consisting of GERAN, UTRAN, E-UTRAN.

23. Method of load balancing according to claim 2, wherein the load related performance indicator being any from the group consisting of: number of redirects to or from a radio access network, number of idle mode UEs in a radio access network, number of periodic Traffic Area Updates (TAU) and load statistics in terms of Classifier Index.

24. Method of load balancing according to claim 22, further comprising, - set, as a mapping table update policy, the priority of an access layer to lowest in a priority order list of at least one entry of the mapping table when load of this access layer is above a threshold.

25. Method of load balancing according to claim 24, further comprising,

- set, as a mapping table update policy, the priority of an access layer one level up in a priority order list of at least one entry of the mapping table when number of redirections to the layer is above a threshold.

26. Method of load balancing according to claim 24, further comprising,

27 Method of load balancing according to claim 24, further comprising,

28. Method of load balancing according to any of claims 24-27, further comprising,

- prioritize a mapping table update policy such that when two or more policies are true only perform the one with highest priority.

29. Method of load balancing according to claim 1, further comprising

- define a total load measure as Total Load =

NO_RFSPl*W_RFSPl+NO_RFSP2*W_RFSP2+...+NO_RFSPx*W_RFSPx, wherein N0_RFSPx denotes number of User Equipments in an access layer having a Classifier Index RFSPx and W_RFSPx denotes an average load of all User Equipments having a Classifier Index RFSPx in the access layer.

30. Method of load balancing according to claim 29, further comprising

- report a Total Load measure to a central node of the mobile communications system.

31. Method of load balancing according to any of claims 29-30, further comprising

- change mapping table in regard of a threshold for a Total Load measure .

32. Computer program, characterized in code means, which when run in a computer causes the computer to execute the method according to any of the claims 1-32.

33. Computer program product including a computer readable medium and a computer program according to claim 32, characterized in that said computer program is included in the computer readable medium.

34. Computer program product according to claim 33, wherein said computer readable medium consists of one or more from the group: ROM (Read-Only Memory) , PROM (Programmable Read-Only Memory) , EPROM (Erasable PROM) , Flash memory, EEPROM (Electrically Erasable PROM), hard disk drive.